I'm not a huge IRV fan or anything but I don't find rangevoting.org to be all that convincing from a US perspective. Most of their references and stats are two decades old or non-existent (e.g. no reference for 80-95% of AUS voters use the NES strategy). Their primary real world evidence is from Australia and Ireland, where independents and third parties currently make up 17% and 47%(!!!) of their parliaments. In the US that number is 0.3% and effectively 0% given how closely Bernie Sanders and Angus King caucus with dems.
Range voting may well be much better, and there are certainly more mathematically sound versions of ranked-choice than IRV, but I think they utterly fail to convince that IRV is just as bad as plurality. They also seem to only take their game theory as far as necessary to reflect Range Voting in the best possible light. For instance, they argue that voters will almost always rank their less preferred of the front-runners last even if they have greater opposition to other candidates, but they don't explore that candidates can and do chase higher rankings among voters that won't rank them #1. It's an obvious and common strategy (candidates were already doing it in my counties first ever RCV election) so I can only assume the reason its not mentioned is that it improves the soundness of RCV in practice.
Their link is referring to the Irish presidential election, which does use IRV—but it’s a meaningless figurehead position, so it’s unclear how relevant the comparison is.
That's a good point, I was grouping IRV and PR-STV when proportional representation isn't a guaranteed component of a ranked-choice system (though many of the dem implemented RCV systems do use it for things like county board or city council seats). Australia's House does use IRV and is at 12% (or 15% if you subtract two vacancies from the major parties).
I am a huge fan of proportional representation/multimember districts, but I think there are some valid arguments that they are not constitutional (and a lot of invalid arguments that may nonetheless carry the day—c’est la vie américaine).
other than the single-winner/multiple-winner focus problem: sure the empirics may be a bit old but there's also the theory? We don't see any reason why we'd magically get better third party support in simulations as well?
Range voting may well be much better, and there are certainly more mathematically sound versions of ranked-choice than IRV, but I think they utterly fail to convince that IRV is just as bad as plurality. They also seem to only take their game theory as far as necessary to reflect Range Voting in the best possible light. For instance, they argue that voters will almost always rank their less preferred of the front-runners last even if they have greater opposition to other candidates, but they don't explore that candidates can and do chase higher rankings among voters that won't rank them #1. It's an obvious and common strategy (candidates were already doing it in my counties first ever RCV election) so I can only assume the reason its not mentioned is that it improves the soundness of RCV in practice.